Human behavior must be flexible, since our environment is continually changing. In the same way that we need flexibility in our social behavior, we also need to adapt to changes in our sensory environment, as in each case rigidity poses a great disadvantage to our survival. A simple example of the latter is the adaptation of the visual system to changes in illumination that occur throughout the course of a day. So efficient is the rescaling of our dynamic range of light perception, that we hardly notice the orders of magnitude difference between noon and sunset. Without flexibility in our visual system, we would be specialists for seeing during one or the other time of day, but not both, and we would spend half of our time nearly blind. Other forms of adaptation can just as well occur over long time scales, such as weeks and months. For example, good evidence suggests that visual expertise, such as that associated with learning new faces, or recognizing brands of automobiles, involves the designation of particular neurons in the brain to specialize for the task. When the associated portions of the brain are disrupted, as in the case of clinical or developmental prosopagnosia, the capacity to discriminate and recognize individual faces can be lost entirely. Patients with this affliction are at a great disadvantage, being unable to recognize even their spouse or own possessions. In our laboratory, we are studying the adaptation of visual processing over multiple time scales, using both psychophysical procedures in humans and electrophysiological procedures in monkeys. In one recently published study, we explored the role of adaptation on the neural processing of shape. In that study, we presented convex or concave patterns for several seconds before probing the selectivity of neurons in cortical area V4 of the macaque brain. We found that neurons changed their analysis of shape even following a relatively brief adaptation period. While the link between these changes and the perception of shape is not yet clear, the nature of the adapation effects fit the predictions of a recently described visual aftereffect illusion, in which a similar period of exposure to complex stimuli affected the manner in which subsequent shapes were perceived. In another study, we characterized the manner in which stimuli change their appearance when viewed for long periods of time. This form of adaptation, which normally escapes perception, was measured using a novel paradigm in which the stimulus itself was changed very slowly to see if an observer would notice. The results suggest that these online changes are not random, but operate according to our internal representation of shape. In particular, they suggest that basic stimulus properties, such as orientation and curvature, are encoded in the brain in a relative, rather than absolute fashion, using certain learned visual "norms" as reference points. These norms are thought to become internalized in the brain through years of experience, and as such represent one way in which the brain adapts its function to its sensory environment. We have previously published electrophysiological findings supporting the existence of norm-based coding in the brain, and further such studies are underway. In this process, it is thought that the adult brain is able to modify large numbers of cells according to the stimuli that it sees. For humans and other primates, some of the most salient and important learned stimuli are the faces of other individuals. The discovery of face cells, or neurons specialized in their responses to one or a small number of faces, illustrates how tailored responses in the brain can become. In the laboratory, we are studying how such neural specialization comes about. While we believe that such neural responses are a product of experience, this has never been proven, since it has previously been impossible to monitor single neurons over days and weeks. Our approach to this problem involves an array of microwires that permits such monitoring, as recently demonstrated in a paper presently under review. In that paper, we found that in the absence of specific learning pressure, the responses of these cells are remarkably stable over time. Our aim now is to train monkeys to discriminate a set of novel faces while monitoring the selectivity of neurons recorded over days and weeks. This approach will answer the important question whether individual neurons permanently change their response selectivity as new faces are learned. Beyond asking whether or not such changes occur, we would further like to understand the principles underlying any such changes. For this we wil adopt the norm-based framework, which we have studied extensively in the past. In particular, we would like to test the hypothesis that as new stimuli are learned within a category (e.g. faces), that the brain automatically extracts the central tendency of their structure, and that this central tendency ("norm") plays a special role in their neural representation. The long-term goal is to determine whether or not such a process may underlie how the adult brain modifies itself in realistic learning situations, as well as in response to injury.